Perch vs Ellipsis pricing
Model token cost plus a 10% platform fee; compute is additionally metered. No seat fee.
What does the review bill depend on?
Model token cost plus a 10% platform fee; compute is additionally metered. No seat fee. Compute is $0.142/vCPU-hour and $0.024/GB-hour. Tokens and execution duration determine the bill; no flat review price is assumed. Ellipsis price source.
| Developers | Perch | Ellipsis |
|---|---|---|
| 3 | About $50/month estimated usage | Needs actual usage and plan entitlements |
| 10 | About $50/month estimated usage | Needs actual usage and plan entitlements |
Perch uses the rounded estimate of about 50¢ per mid-sized PR, approximately 150 methods read at the captured September–October 2026 density of 34 questions per method. At 10¢ per 1,000 fresh checks the unrounded example is 51¢ per PR. Scope, cache hits and repeated reviews change usage. A vendor subscription floor excludes any additional charges and does not guarantee that every review is included. Credit packs, diff kilobytes and reviewed files cannot be converted into PR totals without their usage assumptions. Perch estimate inputs.
Compare the bill for your workflow.
Model token cost plus a 10% platform fee; compute is additionally metered. No seat fee.
Price the checks your team needs
Compute is $0.142/vCPU-hour and $0.024/GB-hour. Tokens and execution duration determine the bill; no flat review price is assumed.
Perch bills fresh rule checks at 10¢ per 1,000, with cached checks free. About 50¢ per mid-sized PR is the September–October 2026 estimate for approximately 150 methods read and 34 questions per method; the unrounded arithmetic is 51¢. Actual scope and cache hits change the bill. New accounts receive $5 free usage credit.
Price sources
Reviewed 2026-10-14. Published by Perch. Prices and allowances can change.